From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-6.8 required=3.0 tests=HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_PATCH,MAILING_LIST_MULTI,SIGNED_OFF_BY,SPF_PASS autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 67A2BECE561 for ; Fri, 21 Sep 2018 01:30:44 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 096E4206B5 for ; Fri, 21 Sep 2018 01:30:43 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 096E4206B5 Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=huawei.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-kernel-owner@vger.kernel.org Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S2388721AbeIUHRC (ORCPT ); Fri, 21 Sep 2018 03:17:02 -0400 Received: from szxga04-in.huawei.com ([45.249.212.190]:13085 "EHLO huawei.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1725927AbeIUHRC (ORCPT ); Fri, 21 Sep 2018 03:17:02 -0400 Received: from DGGEMS403-HUB.china.huawei.com (unknown [172.30.72.59]) by Forcepoint Email with ESMTP id 50CD0327DC35B; Fri, 21 Sep 2018 09:30:38 +0800 (CST) Received: from [127.0.0.1] (10.134.22.195) by DGGEMS403-HUB.china.huawei.com (10.3.19.203) with Microsoft SMTP Server id 14.3.399.0; Fri, 21 Sep 2018 09:30:37 +0800 Subject: Re: [f2fs-dev] [PATCH 2/2 v2] f2fs: avoid f2fs_bug_on if f2fs_get_meta_page_nofail got EIO To: Jaegeuk Kim CC: , References: <20180918021805.85032-1-jaegeuk@kernel.org> <20180918021805.85032-2-jaegeuk@kernel.org> <20180918175648.GE91945@jaegeuk-macbookpro.roam.corp.google.com> <9813f294-3e09-ee2d-56f0-6bb476e88a5a@huawei.com> <20180920214656.GE35918@jaegeuk-macbookpro.roam.corp.google.com> From: Chao Yu Message-ID: Date: Fri, 21 Sep 2018 09:30:36 +0800 User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:52.0) Gecko/20100101 Thunderbird/52.9.1 MIME-Version: 1.0 In-Reply-To: <20180920214656.GE35918@jaegeuk-macbookpro.roam.corp.google.com> Content-Type: text/plain; charset="windows-1252" Content-Language: en-US Content-Transfer-Encoding: 7bit X-Originating-IP: [10.134.22.195] X-CFilter-Loop: Reflected Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 2018/9/21 5:46, Jaegeuk Kim wrote: > On 09/20, Chao Yu wrote: >> On 2018/9/19 1:56, Jaegeuk Kim wrote: >>> This patch avoids BUG_ON when f2fs_get_meta_page_nofail got EIO during >>> xfstests/generic/475. >>> >>> Signed-off-by: Jaegeuk Kim >>> --- >>> Change log from v1: >>> - propagate errno >>> - drop sum_pages >>> >>> fs/f2fs/checkpoint.c | 10 +++++----- >>> fs/f2fs/f2fs.h | 2 +- >>> fs/f2fs/gc.c | 9 +++++++++ >>> fs/f2fs/node.c | 32 ++++++++++++++++++++++++-------- >>> fs/f2fs/recovery.c | 2 ++ >>> fs/f2fs/segment.c | 3 +++ >>> 6 files changed, 44 insertions(+), 14 deletions(-) >>> >>> diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c >>> index 01e0d8f5bbbe..1ddf332ce4b2 100644 >>> --- a/fs/f2fs/checkpoint.c >>> +++ b/fs/f2fs/checkpoint.c >>> @@ -119,11 +119,8 @@ struct page *f2fs_get_meta_page_nofail(struct f2fs_sb_info *sbi, pgoff_t index) >>> if (PTR_ERR(page) == -EIO && >>> ++count <= DEFAULT_RETRY_IO_COUNT) >>> goto retry; >>> - >>> f2fs_stop_checkpoint(sbi, false); >>> - f2fs_bug_on(sbi, 1); >>> } >>> - >>> return page; >>> } >>> >>> @@ -1512,7 +1509,10 @@ int f2fs_write_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc) >>> ckpt->checkpoint_ver = cpu_to_le64(++ckpt_ver); >>> >>> /* write cached NAT/SIT entries to NAT/SIT area */ >>> - f2fs_flush_nat_entries(sbi, cpc); >>> + err = f2fs_flush_nat_entries(sbi, cpc); >>> + if (err) >>> + goto stop; >> >> To make sure, if partial NAT pages become dirty, flush them back to device >> outside checkpoint() is not harmful, right? > > I think so. Oh, one more case, now we use NAT #0, if partial NAT pages were writebacked, then we start to use NAT #1, later, in another checkpoint(), we update those NAT pages again, we will write them to NAT #0, which is belong to last valid checkpoint(), it's may cause corrupted checkpoint in scenario of SPO. So it's harmfull, right? Thanks, > >> >>> + >>> f2fs_flush_sit_entries(sbi, cpc); >>> >>> /* unlock all the fs_lock[] in do_checkpoint() */ >>> @@ -1521,7 +1521,7 @@ int f2fs_write_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc) >>> f2fs_release_discard_addrs(sbi); >>> else >>> f2fs_clear_prefree_segments(sbi, cpc); >>> - >>> +stop: >>> unblock_operations(sbi); >>> stat_inc_cp_count(sbi->stat_info); >>> >>> diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h >>> index a0c868127a7c..29021dbc8f2a 100644 >>> --- a/fs/f2fs/f2fs.h >>> +++ b/fs/f2fs/f2fs.h >>> @@ -2895,7 +2895,7 @@ int f2fs_recover_xattr_data(struct inode *inode, struct page *page); >>> int f2fs_recover_inode_page(struct f2fs_sb_info *sbi, struct page *page); >>> int f2fs_restore_node_summary(struct f2fs_sb_info *sbi, >>> unsigned int segno, struct f2fs_summary_block *sum); >>> -void f2fs_flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc); >>> +int f2fs_flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc); >>> int f2fs_build_node_manager(struct f2fs_sb_info *sbi); >>> void f2fs_destroy_node_manager(struct f2fs_sb_info *sbi); >>> int __init f2fs_create_node_manager_caches(void); >>> diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c >>> index 4bcc8a59fdef..c7d14282dbf4 100644 >>> --- a/fs/f2fs/gc.c >>> +++ b/fs/f2fs/gc.c >>> @@ -1070,6 +1070,15 @@ static int do_garbage_collect(struct f2fs_sb_info *sbi, >>> /* reference all summary page */ >>> while (segno < end_segno) { >>> sum_page = f2fs_get_sum_page(sbi, segno++); >>> + if (IS_ERR(sum_page)) { >>> + end_segno = segno - 1; >>> + for (segno = start_segno; segno < end_segno; segno++) { >>> + sum_page = find_get_page(META_MAPPING(sbi), >>> + GET_SUM_BLOCK(sbi, segno)); >>> + f2fs_put_page(sum_page, 0); >> >> find_get_page() will add one more reference on page, so we need to call >> f2fs_put_page(sum_page, 0) twice. > > Done. > >> >> Thanks, >> >>> + } >>> + return PTR_ERR(sum_page); >>> + } >>> unlock_page(sum_page); >>> } >>> >>> diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c >>> index fa2381c0bc47..79b6fee354f7 100644 >>> --- a/fs/f2fs/node.c >>> +++ b/fs/f2fs/node.c >>> @@ -126,6 +126,8 @@ static struct page *get_next_nat_page(struct f2fs_sb_info *sbi, nid_t nid) >>> >>> /* get current nat block page with lock */ >>> src_page = get_current_nat_page(sbi, nid); >>> + if (IS_ERR(src_page)) >>> + return src_page; >>> dst_page = f2fs_grab_meta_page(sbi, dst_off); >>> f2fs_bug_on(sbi, PageDirty(src_page)); >>> >>> @@ -2265,15 +2267,19 @@ static int __f2fs_build_free_nids(struct f2fs_sb_info *sbi, >>> nm_i->nat_block_bitmap)) { >>> struct page *page = get_current_nat_page(sbi, nid); >>> >>> - ret = scan_nat_page(sbi, page, nid); >>> - f2fs_put_page(page, 1); >>> + if (IS_ERR(page)) { >>> + ret = PTR_ERR(page); >>> + } else { >>> + ret = scan_nat_page(sbi, page, nid); >>> + f2fs_put_page(page, 1); >>> + } >>> >>> if (ret) { >>> up_read(&nm_i->nat_tree_lock); >>> f2fs_bug_on(sbi, !mount); >>> f2fs_msg(sbi->sb, KERN_ERR, >>> "NAT is corrupt, run fsck to fix it"); >>> - return -EINVAL; >>> + return ret; >>> } >>> } >>> >>> @@ -2700,7 +2706,7 @@ static void __update_nat_bits(struct f2fs_sb_info *sbi, nid_t start_nid, >>> __clear_bit_le(nat_index, nm_i->full_nat_bits); >>> } >>> >>> -static void __flush_nat_entry_set(struct f2fs_sb_info *sbi, >>> +static int __flush_nat_entry_set(struct f2fs_sb_info *sbi, >>> struct nat_entry_set *set, struct cp_control *cpc) >>> { >>> struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA); >>> @@ -2724,6 +2730,9 @@ static void __flush_nat_entry_set(struct f2fs_sb_info *sbi, >>> down_write(&curseg->journal_rwsem); >>> } else { >>> page = get_next_nat_page(sbi, start_nid); >>> + if (IS_ERR(page)) >>> + return PTR_ERR(page); >>> + >>> nat_blk = page_address(page); >>> f2fs_bug_on(sbi, !nat_blk); >>> } >>> @@ -2769,12 +2778,13 @@ static void __flush_nat_entry_set(struct f2fs_sb_info *sbi, >>> radix_tree_delete(&NM_I(sbi)->nat_set_root, set->set); >>> kmem_cache_free(nat_entry_set_slab, set); >>> } >>> + return 0; >>> } >>> >>> /* >>> * This function is called during the checkpointing process. >>> */ >>> -void f2fs_flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc) >>> +int f2fs_flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc) >>> { >>> struct f2fs_nm_info *nm_i = NM_I(sbi); >>> struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA); >>> @@ -2784,6 +2794,7 @@ void f2fs_flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc) >>> unsigned int found; >>> nid_t set_idx = 0; >>> LIST_HEAD(sets); >>> + int err = 0; >>> >>> /* during unmount, let's flush nat_bits before checking dirty_nat_cnt */ >>> if (enabled_nat_bits(sbi, cpc)) { >>> @@ -2793,7 +2804,7 @@ void f2fs_flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc) >>> } >>> >>> if (!nm_i->dirty_nat_cnt) >>> - return; >>> + return 0; >>> >>> down_write(&nm_i->nat_tree_lock); >>> >>> @@ -2816,11 +2827,16 @@ void f2fs_flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc) >>> } >>> >>> /* flush dirty nats in nat entry set */ >>> - list_for_each_entry_safe(set, tmp, &sets, set_list) >>> - __flush_nat_entry_set(sbi, set, cpc); >>> + list_for_each_entry_safe(set, tmp, &sets, set_list) { >>> + err = __flush_nat_entry_set(sbi, set, cpc); >>> + if (err) >>> + break; >>> + } >>> >>> up_write(&nm_i->nat_tree_lock); >>> /* Allow dirty nats by node block allocation in write_begin */ >>> + >>> + return err; >>> } >>> >>> static int __get_nat_bitmaps(struct f2fs_sb_info *sbi) >>> diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c >>> index 56d34193a74b..ba678efe7cad 100644 >>> --- a/fs/f2fs/recovery.c >>> +++ b/fs/f2fs/recovery.c >>> @@ -355,6 +355,8 @@ static int check_index_in_prev_nodes(struct f2fs_sb_info *sbi, >>> } >>> >>> sum_page = f2fs_get_sum_page(sbi, segno); >>> + if (IS_ERR(sum_page)) >>> + return PTR_ERR(sum_page); >>> sum_node = (struct f2fs_summary_block *)page_address(sum_page); >>> sum = sum_node->entries[blkoff]; >>> f2fs_put_page(sum_page, 1); >>> diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c >>> index aa96a371aaf8..ab7386a7e9d3 100644 >>> --- a/fs/f2fs/segment.c >>> +++ b/fs/f2fs/segment.c >>> @@ -2487,6 +2487,7 @@ static void change_curseg(struct f2fs_sb_info *sbi, int type) >>> __next_free_blkoff(sbi, curseg, 0); >>> >>> sum_page = f2fs_get_sum_page(sbi, new_segno); >>> + f2fs_bug_on(sbi, IS_ERR(sum_page)); >>> sum_node = (struct f2fs_summary_block *)page_address(sum_page); >>> memcpy(curseg->sum_blk, sum_node, SUM_ENTRY_SIZE); >>> f2fs_put_page(sum_page, 1); >>> @@ -3971,6 +3972,8 @@ static int build_sit_entries(struct f2fs_sb_info *sbi) >>> >>> se = &sit_i->sentries[start]; >>> page = get_current_sit_page(sbi, start); >>> + if (IS_ERR(page)) >>> + return PTR_ERR(page); >>> sit_blk = (struct f2fs_sit_block *)page_address(page); >>> sit = sit_blk->entries[SIT_ENTRY_OFFSET(sit_i, start)]; >>> f2fs_put_page(page, 1); >>> > > . >